"Guidelines for preparation and execution of studies of the social and economic

impact of HIV/AIDS."

Mid-Conference Summary, November 9, 1999

                      Moderator Lori Bollinger has prepared the following summary of the success and proceedings of the Impact conference to date.  At this time we also encourage you to read a summary and download a document provided by UNAIDS for review over the next two weeks entitled

                      Mid-Conference Summary by Dr. Lori Bollinger

                      First, I would like to thank all of the people who have participated in the on-line discussion so far. Everyone is very busy (working on the impact of HIV/AIDS!), so thank you for taking the time to share your information and expertise with the rest of the community.

                      There are three threads in the on-line discussion that I would like to summarize: the research questions that have been raised in the macroeconomic discussion; the data issues that have been discussed; and finally, the need for a data repository.

                      Macroeconomic research issues: There seems to be a consensus that accurate effects at the macroeconomic level are difficult to ascertain. Various people made suggestions as to how estimates could be improved:

  1.               Include a rural focus, with an examination of the impact on food production, and the eventual impact of reduced food supply to urban areas

2.               Include an examination of the different levels of labor productivity in the country, including whether the country has primarily labor-intensive or capital-intensive industries and technology, and the overall quality of labor

 3.               Further develop a model for examining the impact of HIV/AIDS using an endogenous growth framework. The argument here is that the majority of the  impact will be seen as the epidemic proceeds, and so a feedback of the effect is  needed in the model

 4.               Continue and expand the empirical cross-country analysis of effects as in the Bloom/Mahal Black Death study

 5.               Ensure that the projected effects are placed in some kind of context, relative to other priorities the government might have, either health-related or otherwise

                      A forthcoming study described by Keith Jefferis and Robert Greener about the impact of  HIV/AIDS in Botswana incorporated some of the above concerns, including an examination of issues in rural areas, and a consideration of the impact of prevailing capital/labor ratios on future economic performance. The study will also examine the impact of HIV/AIDS on  household-level issues of poverty and economic inequality.

                      Data issues: The discussion has focused on two sets of data issues, availability and quality. This discussion has generated ideas for researchers regarding places to look for data on both

 levels and projections of prevalence; for example, UNAIDS, US Census, and national surveillance systems (by the way, I would still like to know how the World Bank did the cover of "Intensifying Action Against HIV/AIDS in Africa"). Various thoughtful comments about data  quality issues were also raised:

  1.               The reliability of national surveillance site data

 2.               The difficulty of correcting for the female selectivity bias in antenatal clinic data,  to arrive at an appropriate female/male ratio

 3.               The difficulty of correcting for the selectivity bias present in antenatal clinic data  from the fertility-reducing effect of HIV

4.               Whether using AIDS cases reported is a better indicator of HIV prevalence than using surveillance site data

 5.               The validity of examining the impact of HIV/AIDS through changes in mortality rates

                      Suggestions were made by fellow researchers as to how to begin to deal with some of these  issues: using small surveys designed to evaluate the female/male ratio; evidence on the low reporting percentage of AIDS cases, implying that HIV prevalence numbers are more reliable; and pleas for more funding to refine surveillance site data collection and reporting.

                      Data Repository: Elizabeth Pisani and Alan Whiteside both made the point that it is economic data about the impact of HIV/AIDS that is truly lacking, rather than epidemiological  data. Anita Alban responded in part to this issue by discussing a new UNAIDS document that  will establish guidelines for performing socioeconomic studies of the impact of HIV/AIDS, to be available on the UNAIDS web site. Various other people expressed an interest in having a  central repository of data and information about possible effects, similar to the "Cochrane Collaboration." Chester Morris agreed with Anita Alban that papers that might be classified as occupying a "grey area", should be included in such a repository.

                      A first attempt at such a repository is actually part of this on-line discussion. There is a link on  the main discussion page to a brief survey form under "Submit Your Experiences", which can  be used to describe either published work, work in progress, or qualitative findings on programs regarding the impact of HIV/AIDS. The survey form consists of filling out an author’s  name, title of study, date, brief description, and then choosing from possible categories, such as geographic and sectoral coverage. After the database is compiled, it will be a searchable  resource that will remain up at the iaen web site for future reference.

                      Final comment: Finally, I would like to encourage additional postings from the people who

 have signed up for this conference. Are there any other studies available, similar to the  Botswana study discussed by Jefferis and Greener? Either a comment regarding another  study, or an entry into the database, would be very useful for other researchers.

                      I would like to encourage program people, in particular, to post comments. Any expertise or   insight that you have developed while implementing your program could help someone else  who is just beginning an intervention, or who needs help in finding a new direction for an   existing program. This is true for interventions that have occurred at the firm level, or at the   household/community level. Most of the comments that have been posted to date are from economists; speaking as an economist, I know that I need input from field people to design  and evaluate programs. As David Bishai stated so eloquently in his comment on Oct. 21:

                           "Please help the economists of the world figure out how to think about the death

                           of a single Latin American man."